基于频率域分析的姿态估计方法

近年来随着机器视觉、图像处理技术的发展,在许多领域中需要依据所拍摄的图像对其拍摄距离及拍摄姿态进行推测。针对传统姿态估计方法中所存在的嵌入标识会影响图像质量、破坏图像完整性的问题,利用图像的空间频率域分析方法,提出了在图像中嵌入高频信号标识,并观察其空间频率变化,从而推断其姿态变化的方法。实验结果证明该方法在不破坏图像完整性的前提下,有效估计出样本图像在三个轴方向上的旋转角度,从而证实了此方法在姿态估计方面的应用可靠、有效。

Pose Estimation Based on Frequency Analysis

In recent years, with the development of machine vision and image processing technology, it has become a very important task to estimating the relative position and posture between the camera and object in many fields. Using the spatial frequency analysis method, the current pose can be estimated by observing the spatial frequency change of the marker that is embedded in the image in advance. The experimental results show that the method is reliable and effective.